OBJECTIVES The primary objectives were to compare the incidences of severe thrombocytopenia, critical thrombocytopenia and a relative decline in platelets from baseline (≥50% decline) between patients receiving linezolid and those receiving vancomycin. The secondary objective was to assess the relationship between vancomycin trough concentration and adverse platelet outcomes. METHODS A matched cohort study was performed at the Upstate New York Veterans' Affairs Healthcare Network from January 2005 until February 2008. Eligibility criteria were: (i) receipt of linezolid or vancomycin therapy for ≥48 h; (ii) initiation of therapy as an inpatient; and (iii) baseline platelets available for evaluation. Patients who received linezolid were matched 1:1 to patients who received vancomycin. Cumulative incidences and times to event for (i) platelet count ≤50,000 cells/mm(3), (ii) platelet count ≤20,000 cells/mm(3) and (iii) ≥50% decline in platelets from baseline were evaluated. Multivariate analyses were performed. RESULTS The study included 502 patients (251 matched pairs). The occurrences of platelet counts ≤50,000 cells/mm(3) and ≤20,000 cells/mm(3) did not differ significantly between linezolid and vancomycin patients. A ≥50% decline in platelets from baseline was observed in 78 (31.1%) patients receiving vancomycin and 43 (17.1%) patients receiving linezolid (risk ratio 0.55; 95% CI 0.40-0.77). A clear exposure-response relationship was observed between vancomycin trough concentration and ≥50% decline in platelets from baseline. CONCLUSIONS The incidence of thrombocytopenia was low and did not differ significantly among vancomycin and linezolid patients.